| Title | Comparison Between Three-Level Box-Behnken Design and Three-Level Full Factorial Design for Probabilistic Resource Assessment: A Case Study in Atadei Geothermal Field, Indonesia |
|---|---|
| Authors | Marchel Christian SUPIJO, Heru Berian PRATAMA, SUTOPO |
| Year | 2020 |
| Conference | World Geothermal Congress |
| Keywords | Probabilistic resource assessment, numerical model, experimental design, response surface method, box-behnken design, full factorial design, Atadei geothermal field, monte carlo simulation |
| Abstract | The assessment of resources in the geothermal green field is likely to face some difficulties due to the limited number of data and the existence of data uncertainty that will directly have an impact on the accuracy of the determination of resources. Therefore, it is necessary to do probabilistic resource potential assessments that consider the uncertainty of the data using the experimental design method. In this paper, we will explore a comparative study of a three-level design of Box-Behnken Design (BBD) and Full Factorial Design (FFD) using the Atadei geothermal numeric model to obtain probabilistic resource assessment results. The purpose of this study was to determine how the results of the Experimental Design (ED) and Surface Response Method (RSM) were based on the BBD and the FFD method on the results of probabilistic resource assessments. BBD and FFD are used to construct 27 and 81 trial cases, for each case four parameters such as permeability, porosity, liquid saturation, and location of the feed zone with three levels, namely minimum, most likely, and maximum are used for calculation. The prediction results of numerical simulations of the Atedai field for various cases are used to construct polynomial functions (proxy equations) and then applied to Monte Carlo simulations to generate probabilistic distribution of potential power outputs. Both methods have succeeded in estimating more robust electricity potential which covers the entire range of values of important reservoir parameters. Deep analysis of the proxy equation based on R-square (R2) and error of the regression (S) for both design shows that BBD has more robust results rather than FFD. The probabilistic power capacity of Atadei geothermal field using Monte Carlo based on three-level Box-Behnken Design for 30 years production for P10, P50, and P90 are 11.7 MW, 18.2 MW, and 25.6 MW respectively. |